0

Problem

I am currently reading computer science literature in data privacy. However, I encountered the following sentence that seems confusing to me

In other words, we ask for a refinement of Dalenius's goal: that nothing about an individual should be learnable from a dataset that cannot be learned from the same dataset but with that individual's data removed.

The Ethical Algorithm: The Science of Socially Aware Algorithm Design

I understand "that nothing ... from a dataset", but not sure about the meaning of modifier of dataset. From the context of the text, it seems that the sentence should mean "with and without individual's data makes no difference".

However, I think it is still necessary to fully understand this sentence because it uncovers some design principle of data privacy.

Could someone help me? Thank you in advance!

  • So I should still be able to watch the cable channel that they dropped from the line-up because I still have all the other channels?? – Jim Dec 5 '19 at 6:03
6

nothing about an individual should be learnable from a dataset [A] that cannot be learned from [another dataset B, which is incidentally] the same dataset but with that individual's data removed.

We should derive the same information about individual X from both datasets:

  • The original dataset that contains data about X.
  • The modified dataset that does not contain data about X.
  • 1
    Nice. To address OP's sub-question about the modifer: "that cannot be learned from..." modifies "nothing". – Otomatonium Dec 4 '19 at 19:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.